Damien Broderick:
> >All those salesfolk and skyscrapers are not mainly there to move stuff
> >from factories to your home, or even to design the things the factories
> >make. They perform lots of other useful functions.
>
>What, though, when all is said and done? ...
>My question is, what economic functions must be performed in those
>skyscrapers when and if we have a combination of positional synthesisers
>and human-level AI (not necessarily conscious, but the kind of thing that
>happens due to Moore's curve, by general consensus in 20-40 years.)
>
> >Assuming advanced AI is very different from just assuming early nanotech.
>
>But hang on, you've snipped my key linking statement, where I claimed:
>
> >But of course this assumes bizarrely that *only* replicative nanotech
> >happens in the future. Moore's Law hasn't shut down, the computers are
> >getting better all the time, and nano is feeding into that curve with
> >smaller and faster computers built at molecular scales
>
>So I assume any disagreement involves timelines for these innovations.
>Which is where we came in.
I don't think it is very productive to have a conversation about what the
economy will be like in "the future". The future is very long and large,
and there will be many different economies in different times and places.
Most people don't think you can have any useful discussion about the
future, since so many things can possibly change. If we are going to
buck this convention, we should take care to be as disciplined as
possible. If we are careful, maybe we can say something about early
nanotech, early AI, early uploading, or even early space colonization,
at least when each of these things is among the first big new changes.
But it is much harder to say things about late versions of these things,
or or about early versions of these given that several of these other
changes have already happened.
That all said, if you want to assume that nanotech happens in the next
thirty years, but that computers just improve as they have been, I'd say
this is *not* necessarily an AI scenario. Thirty years of the same sort
of improvement we've had in the last thirty years gives you fast machines,
but not necessarily human level software. That software is what is
required to replace all those people working in skyscrapers today.
Robin Hanson rhanson@gmu.edu http://hanson.gmu.edu
Asst. Prof. Economics, George Mason University
MSN 1D3, Carow Hall, Fairfax VA 22030
703-993-2326 FAX: 703-993-2323
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:02:20 MDT